Record: memories belong in their repo; atomic Copilot-request - #1124
Merged
Conversation
…ow-only PR exemption Three reusable learnings from the Morrison-Lab/rpt Claude-workflow migration (rpt#178/#181/#183): - CLAUDE.md "Encoding reusable feedback": memories go in the repo that owns them, not session-local auto-memory, and you commit them rather than offering to. Reinforces the existing "don't ask" rule with the incident that violated it. - shared/workflow/pr-on-claim.md: the no-unreviewed-pr Stop hook discharges a Copilot review request only when the requested_reviewers POST is the last/atomic command in its Bash call; chaining verification after it keeps the hook warning. - memories/github.md: a workflow-only PR is exempt from news.yaml + version-check (paths-ignore .github/workflows/**); adding any other file trips both, fixed with the no-changelog + no-version-increment labels (which re-trigger on labeled). Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
![claude[bot]](https://avatars.githubusercontent.com/in/1236702?s=60&v=4){kind=small}
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
… github.md section - pr-on-claim.md: split the requested_reviewers example so the POST stands alone and the verification reads follow in a separate block, matching the new sole-command rule (it previously showed the chained anti-pattern). - memories/github.md: drop the workflow-only-PR exemption section. It was misplaced (that file is CLI/API mechanics, not Actions config), redundant with the existing more-careful entry in memories/github-actions.md (lines 459-482), and over-generalized to "all UCD-SERG template repos" -- which that entry's own serocalculator counter-example refutes. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
d-morrison
changed the title
Contributor
|
Claude finished review — View run Already verified in prior rounds and unchanged in this diff, so I won't re-run that check. The diff at the current head is identical to what round 2 reviewed (head Re-confirmed independently this round:
No new findings. The prior verdict from 2026-08-04T02:53:31Z — "Needs minor changes... The only remaining item is updating the PR title/description" — has had that remaining item resolved, so nothing stands in the way of merge now. VerdictReady for merge |
Contributor
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Two reusable learnings from the
Morrison-Lab/rptClaude-workflow migration (rpt #178 / #181 / #183), committed to their owning repo per the "Encoding reusable feedback" rule.1.
CLAUDE.md— memories belong in their repo; don't offer, commitReinforces the existing "decide on your own --- don't ask" rule with the move it didn't rule out: parking a reusable learning in session-local auto-memory and then offering to upstream it. Directive: "always put memories in the repos where they belong; don't wait for confirmation." Do/Don't pair + incident.
2.
shared/workflow/pr-on-claim.md— run the Copilot-request POST as the sole/last commandThe
no-unreviewed-prStop hook discharges the reviewer-request obligation only when therequested_reviewersPOST is the last command in its Bash call (itsis_erroris the whole call's exit status). Chaininggh pr view --json reviews/gh pr checksafter it makes the request non-last → the hook treats it as ambiguous and keeps warning even on a 200. This isfail-fast's "a combined result cannot attribute a per-step outcome" applied to a review request. (rpt#181: this re-fired the hook across six turns before it was diagnosed.) The pre-existing example block is split in this change so it demonstrates the rule rather than the anti-pattern.Dropped during review
A third addition to
memories/github.md(a workflow-only-PR changelog/version-check exemption note) was removed indfc9b5b: it was redundant with the existing, more careful entry inmemories/github-actions.md(lines 459-482), misplaced (that file is CLI/API mechanics), and over-generalized to "all UCD-SERG template repos" --- refuted by that entry's ownserocalculatorcounter-example.Self-reviewed:
check-new-line-breaksclean, added lines ASCII-only.🤖 Generated with Claude Code